Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/NicoGenti/opencode-orchestrator-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/nicogenti/opencode-orchestrator-kit/research-mode)<a href="https://agentmods.dev/commands/nicogenti/opencode-orchestrator-kit/research-mode"><img src="https://agentmods.dev/badge/commands/nicogenti/opencode-orchestrator-kit/research-mode/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/nicogenti/opencode-orchestrator-kit/research-mode"><img src="https://agentmods.dev/badge/commands/nicogenti/opencode-orchestrator-kit/research-mode.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00017 | $0.03261 |
| Opus 5.5 | $0.00007 | $0.01304 |
| Sonnet 5.5 | $0.00003 | $0.00652 |
| Haiku 4.5 | $0.00002 | $0.00326 |
Grade A, and why
research-mode scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research mode records engineering-metrics events — task type, duration, agent calls, token counts, normalised confidence fields, DCI scores and the behavioral A/B slot label — to .context/research-dataset.jsonl for each completed task while the mode is active. It is OFF by default; it never activates accidentally.
Activation and alternation
Subcommands:
/research-mode on
Activates research mode for this session. Requires explicit user confirmation before enabling.
Exact confirmation phrase: "Research mode changes session behavior. Confirm activation?"
When confirmed, the session enters the research lifecycle: every task is assigned a research_id and a slot when it begins. Slot alternation is deterministic in activation order: first active task = A, next = B, then A, B, … No re-assignment after the fact.
/research-mode off Immediately stops all recording. No partial event is completed or written. The per-session slot state is reset. Subsequent tasks produce no dataset entries until the next explicit activation.
Config path: research.enabled: true in .opencode/comprehension.config.json (user edit = explicit opt-in). Default when neither is set: OFF.
A/B slots are behavioral, not labels
The slot label selects one of two materially different per-task workflows:
- Slot A (control) — run the normal per-task workflow WITHOUT the comprehension gate: no classification step, no comprehension questions, 0 coach calls, no DCI, no calibration question. Record
gate_type="n/a"(slot A never enters the gate layer).skippedstays false — a control slot is not a user skip; there is no comprehension to skip. - Slot B (treatment) — run the task WITH the full comprehension gate: classify NONE/LIGHT/DEEP, run the gate per
agents/orchestrator.md(coach questions, DCI, possible RETRY). Record the resultinggate_type(NONE/LIGHT/DEEP);NONEstill means 0 coach calls and no calibration question.skippedis true only for a user-issuedskip comprehensionduring an active LIGHT/DEEP gate.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 157 lines · 17 tokens per session scan A 5a37d7e28b73
research-mode is a command published in the GitHub repository NicoGenti/opencode-orchestrator-kit (8 stars, last pushed yesterday), licensed MIT. It adds 17 tokens to every session and 3,261 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-27.
Other commands, from other repositories
init
Install the formatters this repository needs, with every command visible before it runs.
discuss
A command that interviews you in five phases and saves a detailed description of a proposed feature or change. EARS is a structured way to write requirements so they can be checked clearly.
dispatcher
Use when you want the orchestrator to pick the next repo to work on across your whole portfolio — it enumerates candidate repos below the confinement root, resolves free/busy from each repo's session.lock lease, ranks the FREE ones by backlog priority × staleness × readiness, recommends the single most worthwhile one…
mach12:gh-pr-read
Read a GitHub pull request's title, body, and all top-level PR conversation comments; optionally locate an HTML-marker comment.
cost-tracker
Track session costs, understand token spend, and get optimization tips.
import
Why Grok has no session import (use /grok:sessions + /grok:resume).